View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Abertay Research Portal I still think it was a banana: memorable ‘lies’ and forgettable ‘truths’ *Kevin Allan and **Fiona Gabbert * Cognitive Electrophysiology and Memory Laboratory School of Psychology College of Life Sciences and Medicine University of Aberdeen Aberdeen, UK, AB24 2UB ** Division of Psychology School of Social and Health Sciences University of Abertay Dundee, UK, DD11HG *Corresponding author:- Email: [email protected] Tel: (+44) (0)1224 273932, Fax: (+44)(0)1224 273426 1 Abstract Interpersonal influences on cognition can distort memory judgements. Two experiments examined the nature of these ‘social’ influences, and whether their persistence is independent of their accuracy. Experiment 1 found that a confederate’s social proximity, as well as the content and the confidence of their utterances, interactively modulate participants’ immediate conformity. Notably, errant confederate statements that ‘lied’ about encoded material had a particularly strong immediate distorting influence on memory judgements. Experiment 2 revealed that these ‘lies’ were also memorable, continuing a day later to impair memory accuracy, while accurate confederate statements failed to produce a corresponding and lasting beneficial effect on memory. These findings suggest that an individual’s ‘informational’ social influence can be selectively heightened when they express misinformation to someone who suspects no deceptive intent. The methods newly introduced here thus allow multiple social and cognitive factors impinging on memory accuracy to be manipulated and examined during realistic, precisely controlled dyadic social interactions. Keywords: social cognition, memory, conformity PsycINFO Classifications:- 2343 Learning and Memory 3040 Social Perception and Cognition 2 Introduction Bartlett (1932) famously emphasised the importance of ‘social’ influences on remembering almost three-quarters of a century ago. However, only recently has there been a revival of interest in the effect that the views and judgements of others can have on the accuracy of an individual’s own memory. Laboratory-based work on eyewitness testimony has produced a number of studies revealing that social influences generate ‘misinformation’ effects, i.e. systematic memory distortions (e.g., Betz, Skowronski & Ostrom, 1996; Gabbert, Memon, Allan & Wright, 2004; Meade & Roediger, 2002; Schneider & Watkins, 1996; Shaw, Garven & Wood, 1997; Walther, Bless, Strack, Rackstraw, Wagner & Werth, 2002; Wright, Self & Justice, 2000). This literature has repeatedly demonstrated that responses in joint recall and recognition tasks are highly malleable, often exhibiting conformity to the suggestions or judgements of others (Schneider & Watkins, 1996; Shaw et al., 1997; Wright et al., 2000). Discussions between participants, following the encoding of material, can also produce significant conformity in responses on subsequent memory tests that they perform alone (e.g., Gabbert, Memon & Allan, 2003; Gabbert, Memon & Wright, in press; Mori, 2003; Wright et al., 2000). Social influences to conform are powerful, in that they can produce higher levels of memory distortion (Gabbert et al., 2004; Meade & Roediger, 2002) than the post-event narratives that are more commonly employed as a vehicle for misinformation. But the influence exerted by one person upon another’s memory judgements can also be modulated by person perception factors. For example, tendencies to conform can be increased (or decreased) 3 by manipulating the perceptions of a pair of participants about the relative knowledge each has of stimuli they encoded together as a dyad (Gabbert et al., in press). Similar effects can be obtained by manipulating the perceived relative competence of each individual (e.g., Kwong See, Hoffman & Wood, 2001), or by manipulating the overt confidence with which individuals make their assertions to each other (e.g., Schneider & Watkins, 1996; Wright et al., 2000). But how these person perception manipulations alter the social influence exerted by one individual upon another is essentially unknown. Furthermore, research has not yet addressed whether such person perception factors can exacerbate, reduce or possibly eliminate the longer-term effects of immediate conformity upon memory. Progress in addressing such issues has been hampered partly by the complexity of the phenomenon itself, due to the inherently dynamic and variable nature of realistic interactions between individuals. During such interactions, it is likely that heterogeneous motivations to conform may be triggered (reviewed recently by Cialdini & Goldstein, 2004; and see Deutsch & Gerard, 1955). Moreover, these motivations to conform may be modulated in their strength by various factors that need not operate with stability or consistency throughout the entire interaction. For example, the quality of information that a person is able to retrieve from their own memory may modulate their willingness to conform. Similarly, motivations to conform during an interaction may be influenced by each actor’s impression of the other’s memory, which also could change as their encounter proceeds. In order to understand the conformity phenomenon further, in terms of the social cognitive factors on which it may depend, the present two experiments 4 introduce a novel method that produces a realistic illusion of social interaction between two individuals, allowing conformity to be robustly generated and subjected to various manipulations. Each participant is asked to study a series of stimuli (in the present two experiments we used pictures depicting household scenes taken from Roediger, Meade & Bergman, 2001). This study phase is undertaken with an experimental confederate who sits next to the participant. The participant is then taken to an experimental booth, and is instructed that the same is happening to the confederate, and that their memory for the recently encoded materials will now be tested. Here, we used a 2-alternative forced-choice recognition task, which the participant thinks they will perform along with the confederate. It is straightforward to define various ‘rules’ or contingencies (including those developed in research on economics and game theory, see Feyr & Fischbacher, 2003) according to which this joint memory task is to be conducted. In the present two experiments, we utilised an order-of-response manipulation whereby the participant was asked always to respond after they heard the confederate’s answer. By these means each participant was led to believe that they and the confederate were sharing their knowledge of the pictures with one another. A key manipulation involves the confederate’s answers, which actually comprise a set of pre-recorded auditory samples held in .WAV files. This provides us with complete trial-by-trial control over the accuracy, content, tone and phrasing of the confederate’s ‘post-event information’ (PEI), which allows their ‘influence’ to be systematically and very flexibly manipulated either within or across participant groups. Moreover, because participants are led to 5 believe that they are having an actual interaction with another person, attributes within interpersonal interactions that are conveyed vocally can easily be manipulated to study their modulatory effect on memory conformity. As a further example, these methods would also allow the confederate’s character traits to be revealed to the participant, or systematically manipulated, at any point prior to, during or after their social interaction. Using these methods, for Experiment 1 we had two specific goals. First and foremost was to determine whether the use of auditory confederate samples produces a compelling illusion of social interaction and, consequently, conformity. Across two groups of participants, we manipulated the social ‘proximity’ (e.g. Latane, 1981) of the source of influence. One group of participants encoded the picture stimuli along with the confederate, and were then tested under the collaborative conditions described above. For comparison, a second group of participants encoded the pictures on their own, and then performed the forced-choice task under identical test conditions, except that the PEI was understood to be the pre-recorded responses of a prior participant. If the illusion of social interaction works as intended, then we should obtain a far stronger effect of PEI on memory judgements in the confederate, compared to the no-confederate, encoding condition (Gabbert et al., 2004; Meade & Roediger, 2002). This would confirm that our method exposes individuals to a potent social influence, over and above the effect of mere exposure to the information conveyed in the PEI samples. Secondly, Experiment 1 also addressed whether these methods can be used to vary and control person perception effects that may modulate the 6 strength of conformity. To achieve this we manipulated our participants’ perception of the confederate’s confidence on a trial by trial basis, utilising pre-recorded samples that expressed PEI with phrasing and tone suggestive either of low or high confidence. Orthogonally, we also manipulated the accurate or errant status of the PEI’s content. In line with the person perception work described above, the confederate’s confidence levels should modulate the conforming influence that they exert. Experiment 1 also allowed any interactions between the effect of confederate confidence, the accuracy of their